Research map / Multi-agent systems and games / Human-AI and goal-conditioned
HERO: Heterogeneous Evidential Robust Object-Level Collaborative Perception
Hao SI, Ehsan Javanmardi, Hanlin Wu, Manabu Tsukada
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add DPR (detection-level pose refinement) (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.661 -> 0.710 | Improved |
| add delay compensation (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.661 -> 0.736 | Improved |
| add both DPR and delay compensation (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.661 -> 0.737 | Improved |
| add DPR on top of delay compensation (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.736 -> 0.737 | No clear effect |
| use max-evidence geometry selection instead of mean-box averaging (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.661 -> 0.737 | Improved |
| use max-evidence selection instead of mean-box averaging (ablation) | HERO · OPV2V-H (clean) | AP70: 0.783 -> 0.810 | Improved |
| use max-evidence selection instead of PoE (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.737 -> 0.737 | No clear effect |
| use max-evidence selection instead of PoE (ablation) | HERO · OPV2V-H (clean) | AP70: 0.784 -> 0.810 | Improved |
| use max-evidence instead of max-score geometry (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.496 -> 0.737 | Improved |
| use max-score instead of max-evidence geometry (ablation) | HERO · OPV2V-H (clean) | AP70: 0.810 -> 0.826 | Mixed |
| use beta-sum instead of score-mean semantic fusion (with reliability) (ablation) | HERO · OPV2V-H (pose noise 0.6, delay 200 ms) | AP70: 0.673 -> 0.737 | Improved |
| add reliability weighting to beta-sum (ablation) | HERO · OPV2V-H (clean) | AP70: 0.804 -> 0.810 | Improved |
Closest papers by meaning
- An Extensible Framework for Open Heterogeneous Collaborative Perception (ICLR 2024 · similarity 0.89)
- UECP: Uncertainty-Enhanced Collaborative Perception (ECCV 2026 · similarity 0.89)
- Multi-agent Collaborative Perception via Motion-aware Robust Communication Network (CVPR 2024 · similarity 0.87)
- RCP-Bench: Benchmarking Robustness for Collaborative Perception Under Diverse Corruptions (CVPR 2025 · similarity 0.87)
- What Makes Good Collaborative Views? Contrastive Mutual Information Maximization for Multi-Agent Perception (AAAI 2024 · similarity 0.87)
- Generative Map Priors for Collaborative BEV Semantic Segmentation (CVPR 2025 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
